DUDLEY KNOX LIBRARY
NAVAL POSTGRADUATE SCHOOL
MONTEREY CA 93943-510-
NAVAL POSTGRADUATE SCHOOL
Monterey, California
THESIS
A 3D SPATIAL CHANNEL MODEL
FOR CELLULAR RADIO
by
Christos Sasiakos
September 2000
Thesis Advisor:
Ramakrishna Janaswamy
Second Reader:
Tri T. Ha
Approved for public release; distribution is unlimited
REPORT DOCUMENTATION PAGE
Form Approved
0188
OMB No. 0704-
Public reporting burden for this collection of information is estimated to average 1 hour per response, including
the time for reviewing instruction, searching existing data sources, gathering and maintaining the data needed, and
completing and reviewing the collection of information. Send comments regarding this burden estimate or any
other aspect of this collection of information, including suggestions for reducing this burden, to Washington
headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite
1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project
(0704-0188) Washington DC 20503.
1. AGENCY USE ONLY (Leave blank)
2. REPORT DATE
September 2000
3. REPORT TYPE AND DATES COVERED
Master's Thesis
4. TITLE AND SUBTITLE: A 3D Spatial Channel model for cellular radio.
6. AUTHOR(S) Christos Sasiakos
5. FUNDING NUMBERS
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
Naval Postgraduate School
Monterey, CA 93943-5000
PERFORMING ORGANIZATION
REPORT NUMBER
9. SPONSORING / MONITORTNG AGENCY NAME(S) AND ADDRESS(ES)
10. SPONSORING / MONITORTNG
AGENCY REPORT NUMBER
11. SUPPLEMENTARY NOTES The views expressed in this thesis are those of the author and do not reflect the official
policy or position of the Department of Defense or the U.S. Government.
12a. DISTRIBUTION / AVAILABILITY STATEMENT
Approved for public release; distribution is unlimited
12b. DISTRIBUTION CODE
13. ABSTRACT (maximum 200 words)
This thesis provides closed form expressions for the angular distribution in azimuth and elevation planes for a geometrically
based single bounce spheroid model. The geometry of the spheroid is defined by the semi-major axis a and the semi-minor axis
b. The other parameter of interest in the model is the distance D between the base station and the mobile station. The latter is
assumed to be at the center of the spheroid. The mobile station is assumed to be the transmitter, while the base station is the
receiver. This thesis investigates the effects of the above parameters on the angular distribution of the received waves.
Important parameters such as the r.m.s angle spread in azimuth and elevation plane are calculated from the p.d.f. expressions
derived. The behaviour of these r.m.s angle spreads versus the ratio a/D or b/D respectively is also investigated.
14. SUBJECT TERMS Spatial channel Model, Joint TOA/AOA pdf.
Azimuth Plane. AOA Marginal pdf in Elevation Plane, r.m.s angle spread
AOA Marginal pdf in
15. NUMBER OF
PAGES 68
16. PRICE CODE
17. SECURITY
CLASSD7ICATION OF
REPORT
Unclassified
18. SECURITY
CLASSD7ICATION OF THIS
PAGE
Unclassified
19. SECURITY
CLASSIFICATION OF
ABSTRACT
Unclassified
20. LIMITATION
OF ABSTRACT
UL
NSN 7540-01-280-5500
Standard Form 298 (Rev. 2-89)
Prescribed by ANSI Std. 239-18
THIS PAGE INTENTIONALLY LEFT BLANK
11
Approved for public release; distribution is unlimited
A 3D SPATIAL CHANNEL MODEL
FOR CELLULAR RADIO
Christos Sasiakos
Lieutenant, Hellenic Navy
B.S.E.E. , Hellenic Naval Academy, 1988
Submitted in partial fulfillment of the
requirements for the degree of
MASTER OF SCIENCE IN ELECTRICAL ENGINEERING
from the
NAVAL POSTGRADUATE SCHOOL
Septemtfepzt
DUDLEY KNOX LIBRARY
NAVAL POSTGRADUATE SCHOOL
MONTEREY CA 93943-5101
ABSTRACT
This thesis provides closed form expressions for the angular distribution in
azimuth and elevation planes for a geometrically based single bounce spheroid model.
The geometry of the spheroid is defined by the semi-major axis a and the semi-minor
axis b. The other parameter of interest in the model is the distance D between the base
station and the mobile station. The latter is assumed to be at the center of the spheroid.
The mobile station is assumed to be the transmitter, while the base station is the receiver.
This thesis investigates the effects of the above parameters on the angular distribution of
the received waves. Important parameters such as the r.m.s angle spread in azimuth and
elevation plane are calculated from the p.d.f. expressions derived. The behaviour of these
r.m.s angle spreads versus the ratio a/D or b/D respectively is also investigated.
THIS PAGE INTENTIONALLY LEFT BLANK
VI
TABLE OF CONTENTS
I. INTRODUCTION 1
A. BACKGROUND 1
B. OBJECTIVE 9
II. PROBLEM FORMULATION 11
A. INTRODUCTION 11
B. DERIVATION OF THE JOINT TOA/AOA P.D.F.S 13
C. DERIVATION OF MARGINAL AOA PROBABILITY DENSITY
FUNCTION 17
D. MARGINAL AOA P.D.F IN AZIMUTH PLANE FOR THE
SPHEROID SPATIAL CHANNEL MODEL 19
E. MARGINAL AOA P.D.F IN AZIMUTH PLANE FOR THE
SPECIAL CASE OF THE SEMI-MINOR AXIS OF THE
SPHEROID GOING ZERO 25
F. DERIVATION OF MARGINAL AOA P.D.F IN AZIMUTH PLANE
THROUGH A SECOND APPROACH 26
G. MARGINAL AOA P.D.F IN ELEVATION PLANE FOR THE
SPHEROID SPATIAL CHANNEL MODEL 29
III. NUMERICAL RESULTS AND DISCUSSION 33
A. PLOTS FOR MARGINAL AOA PDF IN AZIMUTH PLANE 33
B. PLOTS FOR MARGINAL AOA PDF IN ELEVATION PLANE 37
IV. CONCLUSIONS AND RECOMMENDATIONS 45
A. CONCLUSIONS 45
B. RECOMMENDATIONS 46
vii
LIST OF REFERENCES 47
INITIAL DISTRIBUTION LIST 49
Vlll
EXECUTIVE SUMMARY
This thesis provides closed forms which describe the statistics of the angular
distribution in azimuth and elevation planes for geometrically based single bounce spheroid
model. This spatial model can be generated by revolving an ellipse, which lies on the xz
plane of a 3D Cartesian coordinate system, about the x-axis. This ellipse has semi-major
axis a and semi-minor axis b, respectively. The base and the mobile stations are located in
the xy plane with the mobile station being at the center of the spheroid and the base station
at the origin of the coordinate system. It is assumed that the mobile station is transmitting
and the base station is receiving. The base and mobile stations are separated by a distance
D.
The scattering spheroid model is applicable to a macrocell environment in which the
received multipath signals at the base station are originated after a scattering from the
surrounding environment about the mobile station. Assuming that the scatterers are
uniformly distributed with a constant scatterer density function in the volume V as defined
by the geometry of the spheroid, we derive the marginal angle of arrival (AOA) probability
density functions (p.d.f)- in the azimuth and elevation plane.
Using the derived closed forms for both p.d.f.s we generate plots to examine how
the parameters a, b and D affect the angular distribution of the received waves. From the
marginal AOA p.d.f. in the azimuth plane we conclude that:
i\
1. The higher the ratio a/D, the lower the probability for the received
multipath signals at the base station to be confined to a small angular
region centered about the axis which connects the base and the mobile
stations.
2. The angular distribution in the azimuth plane is independent of the length
of the vertical dimension b of the spheroid .
3. Comparing the results of the spheroid model for small values of b with
those of the circular scattering model presented in [Ref. 3], it was
observed that there exists a significant difference between them.
Additionally, the effect of the a/D ratio on the r.m.s angle spread in azimuth plane is
discussed. From the plot results of the r.m.s angle spread in azimuth plane versus a/D,
we conclude that for values of a/D up to 0.5, there exists a linear relation. However, for
higher values of a/D, the relationship changes to a nonlinear curve with a positive
gradient.
Regarding the marginal AOA p.d.f. in the elevation plane, the following
conclusions were drawn from the plots:
1. Similar to the p.d.f in the azimuth plane, the higher the ratio b/D the lower
the probability for the received multipath components at the base station to
be restricted in a small angular spread around the axis, which connects the
base and the mobile stations.
2. The r.m.s angle spread which also depends on the semi major axis a, has
a linear dependence with respect to the b/D ratio.
AKNOWLEDGMENT
The author would like to thank Prof. R.Janaswamy for his support and his
guidance through every step of this thesis, my wife, Maria, for all her patience, and
understanding.
XI
THIS PAGE INTENTIONALLY LEFT BLANK
xn
I. INTRODUCTION
A. BACKGROUND
The outdoor radio environment contains a large number of obstacles, such as
buildings, cars, hills, etc, which obstruct, reflect and scatter the wireless radio signals. As
a result, the transmitted signal arrives at the mobile station or at the base station from
many different angles and times of arrival; this is known as multipath. Another common
phenomenon in cellular systems is the co-channel interference, and it is created because
in a certain geographic region(cluster) there are other subregions(cells) that utilize the
same set of frequencies. Multipath fading and co-channel interference can increase
significantly the bit error rate and they don't allow high bit rate data services in wireless
communications systems. The presence of multipath and the motion of the receiver or the
transmitter causes the received signal envelope to be time-varying and experience a
Doppler shift that depends upon the angle of arrival of the incoming waves. Concepts
such as angular spread, time delay spread, and Doppler spread have to be taken in
consideration during a wireless communication system evaluation. In particular, they
should be taken into consideration in the design of smart antennas used for diversity,
beamforming and emitter localization applications.
The multipath effect alone can be countered by applying antenna spatial diversity
which implements an antenna array consisting of multi-elements. Spatial diversity is an
arrangement whereby signals from multiple elements are combined to result in a signal
that fades less rapidly than the individual element signals. Ideally this can be
accomplished if received signals at the two elements are decorrelated. The degree of the
decorrelation between the two received signals by the two elements can be expressed by
the spatial cross-correlation coefficient which is a function of the joint p.d.f. f^e{
b,0) , where ft represents the complementary angle of 0b in elevation plane.
11
Figure 6. The spheroid scatterer density geometry.
The method for the derivation of the joint TOA/AOA p.d.f. and marginal AOA p.d.f. for
both azimuth and elevation planes is based on the approach presented in [Ref. 3]. In the
Figure 6, it is assumed that scatterers are distributed randomly in the volume V of the
spheroid according to the spatial scatterer density function fxyz(x,y,z). It is further
assumed that the waves arrive directly at the base station after undergoing scattering only
12
once from within the volume V. Multiple scattering of waves is not considered in this
thesis.
Before proceeding with the determination of the p.d.f.s in the azimuth and the
elevation plane we would like to present the formulation for the joint TOA/AOA p.d.f
and marginal AOA p.d.f in the general case. In section B we start with the derivation of
the joint TOA/AOA p.d.f.. In section C, we show formulation for the marginal AOA
p.d.f.s. In the following sections D through G we derive the marginal AOA p.d.f in
azimuth and elevation plane for the spheroidal model.
B. DERIVATION OF THE JOINT TOA/AOA P.D.F.'S
Figure 7, shows a plane wave impinging on a base station after scattering from an
obstacle, represented by a point 'S' in space with the ordered rectangular triples (x, y,z).
13
S,(x,y,z)
Figure 7. The 3D used coordinate system.
The scatterer probability density function fXjytZ (x, y, z) can be also expressed
terms of spherical coordinates systems ( rb,(/)b,6b) as follows. The equations relating the
spherical coordinates to the Cartesian coordinates are:
-V?
1 9
+ y" +z"
(1)
6u = cos'
V*2+j
V"+Z"
(2)
14
^=tan"1(f)
(3)
x = rb sin 0b cos 0bA)=J(rb,>b>0b)fx,yAx'y>z)
x = rb sin 6b cos b
Z = rb cos 6b
(7)
where j(rb,fo>0b)= Q-Mr-ttr = rb s™0b> o$b>@b) w^ respect to spherical coordinates:
.2 •
Ail (rb,fb,Ob) = rb sin0bf (rb sin 6^ cosb,rb cos0fe)
(9)
If we want to express the scatterer probability density function fr $ q Vb^b^b)
with respect to (T,0b,0b), where T denotes the time delay of the multipath component,
we have first to find the relationship between rb and the T. Applying the law of cosines
to the triangle BS,S,MS we get the following equation:
15
rf = rg + D2 - 2rbDcosS (10)
where the cos 8 is called the 'cosine direction' given by:
cos S = sin6b cos(f)b . (11)
The delay corresponding to the total path traveled by the wave is:
r = -^p-= £ \rb +4^ + D2 + 2rbDsinOb cos^ ) (12)
where 'c' is the speed of propagation of waves.
Squaring both sides of the equation (12) and solving for rb gives:
b 2{D sin 0b cos fo>0b] =
drb
dr
2(Dsin^cos0z,-rc)" ,* c\
D2c+r2c3 -2tc3D sin 6b cos
Substituting the equation (15) into (14) we get for the joint TOA/AOA p.d.f.
f (T a. a ,_ D2c+T2c3-2rc2Dsmdbcos0b f ( p2-T2c2 A n) (ifft
Jr^eb Mb^b) ~ 2{Dsm9b.C0S)
Using the last equation (16) and the equations (9) and (13) we can express the joint
TOA/AOA p.d.f. / ' 0{T,(pb,&b) in terms of original scatterer probability density
function fx v _(x, y,z) as follows:
16
(D2-r2c2)2(D2c+T2ci-2tc2Dsindbcos4>b)
r , n n • n ( L> -T C ) (U C+T C -4ZC U Sin tfh COS«>j, 1 . / . „ . . _ . „ \
fz,0b ,eb (?,foA) = sin eb rTT— " a fx.y.z Vb sm 0* cos $, , rb sin ^ sm (pb , rb cos ^ ) •
" o(i'sin&6 cos0j,-arj
(IV)
Note that the equation (17) expresses the joint TOA/AOA p.d.f. / Q [T,y>Z)=\ ■ (18)
[0, otherwise
If we consider this special case, the joint TOA/AOA p.d.f. / . Q [T,0b,0b) can be
written:
f (t 2-T2c2)2(D2c+T2c3-2TC2Dsm9bcosb) f.Q.
Jr,b,0b) , and then either with respect to
0b\b) over the range 0b(b) or with respect >b{0b) over the range
(f)b (0b ) to ^ \0b ) depending on whether the marginal A.O.A p.d.f. is to be found in
17
azimuth or elevation plane respectively. The terms rb(0b,9b) and rb (0b,9b) are the
lower and the upper limits of integration and depend on the azimuth angle b) in azimuth plane is given by:
UMb)=i)A jfrb,M(rb,b,0b)drbd0b =
fr] V^i I
r„Mb^bl
j* rb sin 9b fx,yyZ (rb sin 9b cos 0b , rb sin db sin (j>b , rb cos 9b ]drbdQb . (20)
The marginal AOA p.d.f. fe \0b ) in elevation plane is given by:
(A A)
fe»^ = lyb) \frbAA^rb^bA)drbd(l)b =
b , rb cos 9b )drb db ) in azimuth plane expressed
as:
/*(*») = £ jK(^,£)-r^,£)]cos£# (28)
20
D cos ,
(j)b = cons tan t
■J a2 - D2 sin2 b
Figure 8. The intersected region R between a plane i . .
«" - D" sin" b and the center located at Dcos(j)b . For a general angle /?,
the straight line from the origin intersects the region R at two points r^ and r^ . At the
two boundary points the roots >i and ru of the quadratic equation (26) in rb coincide.
This happens when the discriminant of the quadratic equation is set to zero. Carrying this
out, the upper limit /?max is given by:
21
Anax = tan
-1
r
b \g--D -sirr^
a V D2_fl2
(29)
This equation is valid for b in the range of sin (- -g-J < ^ < sin Hjj. These
correspond to the values of ^ that bound the spheroid in the xy-plane. Taking all of
the above into consideration, the marginal AOA p.d.f. in azimuth plane is given by:
Lift)
f*W>) = W Jcos^^^J-^^^^^in'l-fj^^^sm-1^). (30)
0
The evaluation of the integral as defined in the equation (30) is straightforward.
The following relations simplify some of the terms:
= (^%f^-^(^>^X^(A.^ + r44(A,^(^f^) + ig(^>^)). (31)
%{b,P)-\(b,P)=1fc
(32)
(33)
(34)
(35)
Using equations (31) through (35) in equation (30) we get:
Anax (& ) , .
f*,Wb) = W ]{% -^K + Vb, + rb2Jcos fid/3 =
o
AnaxWfc)
3V J
^B2(/3)cos2^-4A(j3)C Ib2(/3)cos2
MP) \ A2(/3)
C
A(£)
^cosfidfi
(36)
22
At a specific value of b when P = Pm2X the discriminant becomes
A(A„ ) = 5: (^ ) cos2 ^ - 4A(#nax )C = 0 . Hence, wnting
= [£2 (/?)cos2 A - 4 A(^)C] - [£2 (/?max )cos2 fa - 4A(/?max )C] ,
we see that
A(^) = 4^2[a2(a2sin2^ + ^2cos2^)-D2a2sin2^-D2Z?2cos2^sin2^ -
-4b2[a2[a2 sin2 0max +b2 cos2 J3max)-D2a2 sin2 jflmax -Z)V cos2 y#max sin2 ^ j,
(37)
which gives:
A = 4Z72[sin2?mu-sin2^Ia2(D2-a2)+Z?2(a2-D2sin2^)]. (38)
Using this the integral in equation (36) can be written as follows:
r (± \ _ 160V cos2 0b^a-{Dl-a-yb2{a--Dz%m2$b) f cos3 /g-y/si^fl^-siir 0 ,g
J ^max _ . ? o, /~~ 2~o 2 a
r cos ^sin /?max -sin /? _ r cosy3(l-sin yg^sin /?max -sin yg
J (a2sin2y3+62cos2y5)3 J (a2sin2/?+62cos2/?)3
On using the substitution t = sin (3 , and fmax = sin /?max the last integral becomes:
"max _,, . i - f" •> _ '. 7 /» max /, -> \ /~t 7
J (ahm20+b2cos2/3)3 P~ J [^+(fl2_fe2^f
23
Hence the term 1— t can be written as follows:
i-<2=-^l-«24242-62)2)J.
Using this in equation (41) we get:
'max I ■> \ Hj 2 max H> 2 max /"^ 2*
f '-rNU-r,f = ^ [ V^CL -df — -L- f 3fa± 2)rmaxSin2Qrj
The last integrals described in the equations (43) (44) and (45) are evaluated using the
formulas 2.563-1, 2.563-2, and 2.563-3 from [Ref. 5].
The marginal p.d.f. in azimuth is then given by:
U(h)=lM-lM)- sin-'f-^ASsin-'fe) (46)
. / \ ;rD2cos2^Va2(D2-a2K*2(a2-D2sin2^) (2a2-4b2)y2+Lb4-2a2b2}y-a2b4 ,A~,
where lx [b ) = "— J ^ '- J ' ^ - ' (47)
Y^y\a--b-j
and IM„)= ^--^Mo^V-o2™-^ {,-?). (48,
iV i]y{a--b-}
24
with y = a2 sin2 J3max +b2cos2j3T]
E. MARGINAL AOA P.D.F IN AZIMUTH PLANE FOR THE SPECIAL
CASE OF THE SEMI-MINOR AXIS OF THE SPHEROID GOING ZERO
In this section we examine the special case of marginal AOA p.d.f., given by
equation (46), for values of 6— »0. In order to evaluate the p.d.f. for this case the
following simplifications are made on equations (47) and (48):
y = a2 sin2 £raax +b2 cos2 ^max = b2 cos2 /?max (l + £ tan 2 /?max)
where tan /?max is given by equation (29). Since b —» 0 the last equation can be
^2 2
expressed as follows: Y -,cos, • (49)
D —a
1 2/2 2 1
The terms y-b and (y-b )\D -a J can be expressed as :
y-b2 =b2a2-D*sm;*b >0md (y-b2)(D2 -a2)=b2(a 2-Z)2sin2^ J. (50)
Plugging equations (47) and (48) to equation (46) the latter can be written as :
I U)-I U )- £ MP2^2M«2-P2*"2ft)
-(D2-a2)(r-^2)+^^[(3a2-4^2V2 + (4^-2flVV-«V
(51)
The term [3a2 -4b2 jy2 + \4b4 -2a2b jy-a2b4 — can be expressed as:
\3a2 -4b2)y2 + (4b4 -2a2b2)y-a2b4}± = 3a2y-2a2b2 -a2^-4b2y + 4b4. (52)
25
The first three terms in the right side of equation (52) are second order terms, while the
last two are fourth order terms with respect to b therefore can be neglected. Hence
equation (52) reduces to:
3a2y-2a2b2 -a2^--4b2y + 4b4 =r4^
[D'-a'
3D2 cos2
i°2-°2{-
D'-a-
\D2-a2\ + -£=*
D cos
-> \
(53)
Combining equations (49), (50) and (53) and taking also in consideration which terms are
second or fourth order with respect to 'b', equation (51) can be written as follows:
'lk)-'2(A)s#fe&
a2 +D2 sin2 ^
3D' cos' 0
is given by:
UM) = jjj{"2-D2™20b) sin^-fj^^^sin-1^). (56)
F . DERIVATION OF MARGINAL AOA PDF IN AZIMUTH PLANE
THROUGH A SECOND APPROACH
As we have mentioned before the marginal AOA p.d.f in azimuth plane is given
by the:
f,hW = V J jrb2sm0drb0 = ±j \rbrb cos/3drbd/3.
(57)
26
D cos Qh - a cos {n-a)
Dcos(|>
ya2 - D2 sin2 b+A cos a]dx2dy2
-AnaxM^) ° °
(59)
The integral in equation (59) with respect to polar coordinate system A and a can be
expressed as:
f0b (0*) = v J J (D cos 0b + A cos otydMa ■ (60)
ar=0 0
Taking in consideration that the term A cos a is equal to X2, since X2 is an odd function,
equation (60) can be written as:
/*to) = v J J(0cosft,+jc2)fctate = £ jJDcos0bAdAda (61)
a=0 0 ^(^)
where A0 (^ ) is the area of region R.
Equation (61) is simplified as:
f^0b) = iDcos0b jJAdAda = ^A^(0b) (62)
A(&)
where A^ ((/>b) = 7T — [a -D sin"^j. Substituting the above term in equation (62) we
get for the marginal AOA p.d.f. in azimuth plane:
4W = ^M«2-fl2sm2<0 sm-(-f) 0 , is also true for all b.
2.X
Finally, using the closed form expression for the p.d.f. we evaluate the r.m.s
angular spread in closed form:
rms = 2b=ir J k fe , /?) - i (b , /?)]cos 0d*b (66)
^ (P)=fnax
ff(P)=& fK(0)-rfa>0)]p*&*>. \fe0u <67>
0
The maximum elevation angle /?M is obtained from the values that bound the
spheroid in the xz-plane and can be found analytically from the equation (29) by setting
the azimuth angle Qb equals to zero:
■_1' h ^ (68)
/?M=tan-'(74=T).
29
Using the equations (31) through (35), the integral in equation (67) can be simplified as
follows:
fP(P)= w J* cos /% - \ K + \ \ + rl
Vmax I — / \
2cos/? f tJB- {p)cos-0b-4 A(/3)C (b2(/3)cos2 0b g_ U.
' 3V J A(0) \ AriP) A(fi)rrb
0
- I^f£ j WR^IC* • (69)
0
Substituting the functional forms of A{(3) , B{(3) and C from equation (26) we get:
2 4 3/5 "max i
ff{p)=vlvZl°?\tfV J ™?h4B2{PWh-*KP)cdb-
H 3(a sin p+» cos p) V J
-^rfe^v / M^°s2fe-4A(/?)o% . (70)
3(<3 sin B+b cos p) V J
0
The upper limit for the azimuth angle 0max , for a fixed {3 , is the largest value of angle
0£ that bound the region R , which is generated by the intersection of the cone
j3 = constant with the spheroid. In general, the intersection will include all the points
rub -4A(/3)C under the square root.
This is done applying the same approach as in previous section D. The discriminant, at a
specific value of (3 becomes equals to zero when the azimuth angle (j)b coincides with
the maximum value of b - 4A(/3)C] - [B2(/3)cos2 0max - 4A(j3)C] , becomes
A((/>b)= \*D Vcos2 ficos2 b- (74>
3(a sin p+b cos p) V J
The above two integrals may be rewritten as:
0max
J
0
0max
f COS 2 0fc V(sin 2 ^max ~ sin 2 ^ )<% = sin ^max / cos 2 h JO ~2T sin 2 ^ Wi ^5)
_ . V Sln ^max
rraax , -rmax i
and b )db )<% . (76)
_ _ V Sln ^max
These integrals can be evaluated in terms of complete elliptical integrals first and second
31
kind. Indeed with the help of the formulas 2.595-1 and 2.592-2 of [Ref. 5], these
integrals take the form:
Tmax
in0 cos"^.
sin
and
max
^(smC(sin ^max ) (78)
max 0
where AT(sin 0^ ) and £(sin 0max ) denote the complete elliptical integral of the first and
second kind respectively [Ref. 5].
Substituting the results from the equations (77) and (78) into equation (74) we get
the final expression for the marginal AOA p.d.f., in elevation fn\P) in a closed form as
follows:
fp(P) =
3L6 4
16D-Vcos4/?
6(a2sin2/?+62cos2/?)3V
1 + sin 2 0max )£(sin 0max ) - cos2 0max K(sin 0max )
4Dfc4(p2-q2)cos2^
3(a2 sin2 P+b2 cos2 /3)2V
£:(sin^max)-cos2^max^(sin^max)J' \0\ ^ &M (79)
^D2-a~
with fiM = tan"1 T^-? , sin^ =ffi-(JJ^f and V =f ;za2£.
32
III. NUMERICAL RESULTS AND DISCUSSION
Having derived closed form expressions for the marginal AOA p.d.f. in azimuth
and elevation planes, we now present in this chapter some numerical results for each
marginal AOA p.d.f.. Recall that D is the distance between transmitter and receiver, a is
the semi-major axis of the spheroid, and b is the semi-minor-axis of the spheroid.
A. PLOTS FOR MARGINAL AOA PDF IN AZIMUTH PLANE
In this section we consider the cases of 'D', 'a' and 'b' as given in Table 2 and
generate some plots:
D(m)
a(m)
b(m)
Corresponding
Figure
1000
100
0.001
10
1000
100
50
11
1000
200
150
12
Table 2. Parameters of D, a and b used in plots for marginal AOA p.d.f. in the azimuth
plane.
Figure 10 has been plotted in order to check how the azimuth angular distributions
of the spheroid and the circular scattering models are related to each other. The circular
scattering model is presented in [Ref. 3]. The angular distribution for the circular model
is given by equation (80):
33
f&M
•3Dcos»VoW»-D2V,sin-l(-t)b
15
Figure 10. Plot of f^ [tfo) versus^ forD=1000m, a=100m and b=0.001m. Also shows
the AOA p.d.f. for a circular spatial model for D=1000m and a=100m.
In Figure 10, we plot the p.d.f. for a/D=0.l and b/D=W6. Comparing the plot of the
spheroid model for small values of 'b' with this of the circular model, we can observe that
there exists a significant difference between them.
In Figures 1 1 and 12 we plot the p.d.f. for a/D=0.l, £=50m and for a/D=0.2,
6= 150m using the two different formulas for the angular distribution in the azimuth
plane.
34
~*4
pdf (eq(46) for
a/D=0.1 b=50m
— •— pdf eq(56) for
a/D=0.1
i _ _ i. i _ +_
jf j. ~v
f: i |
-15 -10 -5 0 5 10
Angle of Arrival in Azimuth Plane(degrees),b
15
Figure 11. Plot of /,* (b
15
Figure 12. Plot of /^(fo) versus b for D=1000m, a=200m and b=150m
(using the two different formulas).
35
As shown Figure 1 1 and 12, the two different formulas give identical results for the
angular distribution in azimuth plane. Additionally, we conclude that the higher the ratio
a/D, the lower the probability for the received signals at the base station to be restricted in
a small angular region centered about the axis which connects the base and the mobile
stations.
Finally in Figure 13 we plot the r.m.s angle spread in azimuth plane as given in
equation (64) versus the ratio a/D.
60
50 -
40
■S 30
20
10
I I
i i i i i
: : ;
r.m.s angular spread 0
rms
M M M l/i
\ \ ! j/i \ ! I
I i^n } ': i i ':
Jn III!!!]
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
a/D
Figure 13. Plot of the r.m.s angular (jw spread versus a/D.
Figure 13 shows that for values up to 0.5, of the ratio a/D, the profile of the r.m.s angle
spread is almost linear. Note that the r.m.s angle spread in azimuth plane is independent
of b as proved with the derivation of the second formula for the p.d.f in azimuth plane.
36
B. PLOTS FOR MARGINAL AOA PDF IN ELEVATION PLANE
D(m)
a(m)
b(m)
Corresponding
Figure
1000
100
1
14
1000
100
30
15
1000
100
50
16
1000
200
50
17
1000
500
50
18
Table 3. Parameters D, a, and b used in plots for marginal AOA p.d.f. in the
elevation plane.
Figure 14 is generated to demonstrate the angular distribution in elevation plane
for small values of b.
37
800
700
600
500
S 400
300
200
100
I
] I
D=1000m
a=100m
b=1m
I
I
]
I
I
I
I I I
4-3-2-101234
Angle of Arrival in Elevation Plane(degrees),fS
Figure 14. Plot of fp{0) versus ft forD=1000m, a=100m
and b=lm..
As we can see form the Figure 14 the angular distribution in elevation plane for small
values of b looks like a dirac function as it is expected.
Figure 15 and 16 are plotted for demonstration purposes, giving information of
how the changes of the ratio b/D for a given 'a' affect the angular distribution in elevation
plane.
38
30
25
20
S 15
10
1 1 1 1 1
D= 1000m
a=100m
b=30m
r -i
i 1
;
i 1 7 ! i ! \
■
-3-2-1 0 1 2
Angle of Arrival in Elevation Plane(degrees),P
Figure 15. Plot of fp(0) versus j3 forD=1000m, a=100m
andb=30m..
Figure 15 shows the the angular distribution in elevation plane for a ratio b/D=0.03 and
a/D=0.l.
16
10
S 8
II ! I | .... .
i
D= 1000m
a=100m
b=50m
I I
L L
:-----i/---r---H----;----\
1
L /_J J J i
V
:/ i i i i
\:
-3-2-1 0 1 2
Angle of Arrival in Elevation Plane(degrees),f3
Figure 16. Plot offp(0) versus y# forD=1000m, a=100m
and b=50m.
39
Figure 16 shows the p.d.f for a ratio b/D=0.05 and a/D=0.1 As seen, compared to Figures
15 and 16, the higher the ratio b/D the lower the probability to receive multipath
components from a small region /3BW at the base station.
Figure 17 is generated, as set with Figure 16, in order to show the effect of the
length of the semi- major axis a on the angular distribution.
S 8
1 1 1 1 1
i
!____! I-/ _l _ AJ
— D= 1000m
a=200m
b=50m
•4"|n"4"-i\
\ t \ :
7
1 - T
7
/
-3 -2 -1-0 1 2
Angle ol Arrival in Elevation Plane(degrees),fS
Figure 17. Plot of ffi(0) versus j3 forD=1000m, a=200m
and b=50in meters.
Figure 17 shows the p.d.f for a ratio b/D=0.05 and a/D=0.2. Comparing Figures 16 and
17, it is seen that p.d.f in elevation plane is almost independent of the length of semi-
major axis of the spheroid for small values of the ratio a/D. Figure 18, is generated to
show the effect of higher values of the a/D ratio on the angular distribution.
40
14
12
10
S 8
4- 1 1- J + V -A
D=1000m
a=500m
b=50m
: :/ 1 : : \: :
: / \
t ■*/ 1 i t ~\ \\~ r
\. 1 1 1 + -( 1 V(_
-3-2-1012
Angle of Arrival in Elevation Plane(degrees),P
Figure 18. Plot of ffi{fl) versus j3 for D=1000m, a=500m,
and b=50m.
Comparing Figure 17 and 18, it was observed that there exists a difference between the
two angular distributions.
Finally in Figure 19 we plot the r.m.s angular spread of the angle of arrival in
elevation plane, as given by the equation (81), versus the ratio b/D for specific values of
a/D.
firm=2\]fi2ffi(fi)dfi
(81)
The mean value of the angular distribution at the base station is zero, and the integral is
evaluated numerically.
41
30
25
_20
w
CD
B 15
0Q.
10
I I
+
B fora/D=0.1
r rms
3 fora/D=0.2
rms
3_0fora/D=0.5
rms
—
0
0
1 0
2
0.3 0.4 0.5
b/D
Figure 19. Plot of the r.m.s angular (3,-™ spread versus b/D
for various ratios a/D,for the spheroid scattering model.
As shown in Figure 19, there exists a linear relation between the r.m.s angular spread and
the ratio b/D independently of the ratio a/D.
Eventually for validation purposes of the two derived, in this thesis, closed
forms of the angular distributions, we plot in Figure 20 both p.d.f. with a almost equals to
b (b is not made exactly equal to a to avoid dealing with limiting process).
42
A, , ,-.
1 1 I 1
a=100m b=99.999m
pdf in elev for D= 1000m
a=100m b=100m
' __ _.:x^
! A \ ,
_ J _/ 1 _ i 1
: / : : : \
:__/.:_ J l L__\_J
..-X.J--i--i-4--
..-/-J--i.-.i-X„,
.-./i...J..-.L-.i....AH
/ : ; : : : \
/ i i i : : \
Figure 20. Plot of f* (